Micro computed tomography (µCT) is an effective way to obtain rock digital models that can be used, along with digital rock physics (DRP), to calculate rock properties. However, DRP has yielded mixed success that are in part due to limitations and uncertainty introduced by the segmentation processing method. Lately, segmentation-less DRP has been put forward potentially enabling DRP to become an avenue to inexpensively calculate rock properties. Ideally, DRP could be applied to an infinite number of samples such as drill cuttings. Here we present an investigation about effective medium theories that in segmentation-less DRP are the basis to assign properties to digital models. We compared a classic segmentation model with modifications of the Hashin-Shtrikman upper bound, and modifications of the Voigt-Reuss-Hill bound. The elastic property matrices that we obtain with these effective medium theories are used as models to simulate the propagation of ultrasonic waves. All effective medium theories applied to the rock as a whole significantly over predicted P wave velocity from 23% to 75%. Segmentation predicted a velocity 64% higher than the lab measurement. In all scenarios, the segmentation-less method predicts wave speeds more accurately than segmentation or their respective whole rock based DRP. For instance, a modification of the Voigt-Reuss-Hill bound yields a P wave velocity within 2% of that measured in the laboratory, while a modification of the Hashin-Shtrikman bound is within 19%.
Presentation Date: Wednesday, September 27, 2017
Start Time: 3:55 PM
Location: Exhibit Hall C, E-P Station 4
Presentation Type: EPOSTER
